Hockey analytics consulting

I know the game, and I build the tech.

I help hockey clubs and sports organisations turn data into decisions — from game and player analysis to the data platform and AI tools behind it. I am an experienced data and AI consultant for sports organisations and other businesses, and I have spent years on the ice and behind the bench.

Start a conversation

Why work with me

I build all of it myself

Data pipelines, models, dashboards, APIs and AI agents — one person from raw data to the tool the coach uses. No hand-offs between an analytics firm, a developer and an IT department.

I know hockey

I played junior hockey in HV71 and coached at HV71 and Nacka Hockey. I know which questions coaches actually ask, and how a number has to look to be used on a Tuesday practice.

In my day job I build data platforms, machine learning and AI agents in production for a global manufacturer. The same engineering standards — governed data, tested pipelines, measurable effect — go into the hockey work.

What I help with

Game and player analysis

Shot, possession and line analysis, opponent scouting and player development tracking — turned into answers coaches can act on between games, not reports nobody opens.

Data platform for the club

Collecting game, tracking and video data from different sources into one governed model, so analysts, coaches and management work from the same numbers.

AI tools and agents

Natural-language questions over your own data, automated pre-game briefs and alerts when something changes — with people approving what matters.

Recruitment and squad planning

Structured scouting data, comparable player profiles and roster scenarios that support decisions instead of replacing judgment.

Strategy and advisory

Where analytics actually pays off in a hockey organisation, what to build first, and how to make the staff use it.

Examples of analyses

A few of my own example analyses on public league data:

Returning cores beat coaching changes

Mean points percentage by returning-core share

Across 52 SHL team-seasons (2021/22–2024/25), returning-core share correlated with points% (r = 0.32); mean age did not (r = −0.05). Teams that changed coach moved about as much as teams that kept theirs. For summer planning: weigh who stays over the headline hire. Read the method noteData: SHL regular season 2021/22–2024/25, public rosters and standings.

A timeout is a trailing-state signal

Share of timeouts by score state when called

1,303 SHL timeouts in 1,854 games (2020/21–2024/25): 81% came in the third period and 83% were followed by a pulled goalie. Read it as the setup for a late 6-on-5, not a momentum reset. Data: SHL regular season 2020/21–2024/25, official play-by-play.

Pulling the goalie when down by one

Outcome after a late pull, down by one

620 late goalie pulls in the SHL (2020/21–2025/26): the trailing team tied it 16.1% of the time but conceded into the empty net 37.7% — and won 7.6%. Useful for setting expectations and practising the 6-on-5. Data: SHL regular season 2020/21–2025/26, official play-by-play.

Is there goal momentum?

Next goal within three minutes

After any SHL goal (2020/21–2025/26), the next goal came within three minutes 23.1% of the time — what random scoring predicts (23.2%). After a go-ahead goal it was even lower. A useful check before building a game plan around "momentum". Data: SHL regular season 2020/21–2025/26, official play-by-play.

Other kinds of analysis I build for clubs:

Game and possession analysis

Shot locations, shot attempts and possession by game state and special teams — for example how the power play behaves at home versus on the road — summarised as a few win factors the staff can follow week to week.

Player and recruitment profiles

Role-based player profiles that compare production with usage and deployment, so scouting lists are filtered on what a player is asked to do — not on a single ranking number.

Opposition scouting

Automated pre-game briefs on the next opponent: tendencies by period and game state, special-teams patterns and line usage, ready before the video meeting.

Workload and development tracking

Shift length, ice time and game load over a season, and development tracking for junior players — covering practice as well as games.

Notes from LINHAC 2026

At the Linköping Hockey Analytics Conference in June 2026 (36 sessions), the same message kept coming back. My main takeaways:

Data is infrastructure, not an edge. Clubs that treat analytics as a side project fall behind.
Own the data, not the dashboard. The lasting advantage is owning data, models and process — not buying another tool.
Insights are cheap, behaviour change is hard. The work is getting players and staff to act on the numbers; weekly reports beat game-by-game noise.
Cover practice, not just games. Most clubs track games but not practice; the chain from capture to decision should cover both.
Simple research signals pay off. Pulling the goalie earlier when down one, and home ice mattering most on the power play, were among the clearest findings.

Read the full LINHAC 2026 report.

Experience with clubs

I have worked with clubs in HockeyAllsvenskan, a club in Denmark and clubs in Switzerland on analysis and data questions — from game and player analysis to how the data is collected and used by the staff.

Background and writing

Player profile on Elite Prospects — HV71 J18 and J20 SuperElit, 2004–2009
Coaching profile on Elite Prospects — HV71 U16 head coach, HV71 J20 assistant coach
LINHAC 2026 conference report — notes from the Linköping Hockey Analytics Conference
SHL returning cores vs coaching changes — a method note on SHL data

What I build

The same engineering goes into the hockey work. Some of my other builds:

Enterprise data platform — governed lakehouse and ML in production
AI analytics platform on Databricks — questions in plain language turned into validated SQL
AI automation in production — pipelines, flows and guardrails
Agentic data organisation with Claude — agents that build and maintain the data platform
AI feedback agent instead of a survey — conversational data collection
AI mail and calendar agent — zero inbox without reading email

Frequently asked questions

What does a hockey analytics consultant do?

A hockey analytics consultant turns game, player and tracking data into decisions for a club: game and possession analysis, opposition scouting, player and recruitment profiles, and the data platform and tools that keep those answers coming every week.

Who is Emil Ingemar Karlsson as a hockey analyst?

Emil Ingemar Karlsson is a Stockholm-based data and AI consultant and hockey analyst. He played junior hockey in HV71, has coached at HV71 and Nacka Hockey, and builds data platforms, machine learning and AI agents in production.

Can a sports data analyst also build the club’s data platform?

Yes. Emil builds the whole chain himself — data collection, pipelines, models, dashboards, APIs and AI agents — so a club gets the analysis and the platform behind it without hand-offs between several suppliers.

Which kinds of hockey analysis are most useful for a club?

Usually a few repeatable ones: roster continuity and squad planning, game and special-teams analysis, role-based player and recruitment profiles, pre-game opposition briefs and workload tracking for players.

Does Emil work with sports organisations beyond hockey?

Yes. He works as a data and AI consultant for sports organisations and other businesses, on data platforms, analytics and AI automation.

Which leagues has Emil worked with?

He has worked with clubs in HockeyAllsvenskan, a club in Denmark and clubs in Switzerland, and he publishes analysis on SHL data.

How do I start working with Emil on hockey analytics?

Send a message through the contact form at emilingemarkarlsson.com/contact describing what your club or organisation is trying to decide.

Have a question your data should answer?

Tell me what your club or organisation is trying to decide, and I will tell you what it would take. Use the contact form — the message comes straight to me.

Start a conversation